254 E. E. FREE 



series, or, for that matter, between any plants selected one from 

 each series, the resulting curve gives the comparison between 

 the different series and enables the reduction of all to the same 

 basis, conveniently expressed on a scale of 100 for the plant in 

 best condition among them all. Obviously a similar procedure 

 permits the reduction of all comparisons to terms of any single 

 standard which may be chosen, as, for instance, the condition 

 of a control plant. In like wise it is possible to reduce the results 

 of a series .of comparisons to actual quantitative terms provided 

 the criteria of judgment be such as to be measurable at all. 

 Suppose, for instance, that plants have been compared as to 

 their relative leafiness. If then the leanness of any two of the 

 plants be actually measured by counting the leaves per meter 

 of stem or in any other way, the insertion of these measurements 

 in the curve resulting from the original comparisons gives a 

 scale by means of which it is possible to reduce the whole com- 

 parison curve to approximately quantitative terms. Procedures 

 for the use of photographs for the comparison of series not 

 synchronous or in proximity and for many other detailed modifica- 

 tions and extensions of the method will suggest themselves at once. 

 One caution is important, especially in connection with the 

 method of weighted scores. In using this method on a large 

 number of individuals — in one investigation the writer employed 

 it with nearly two hundred — the mental criteria for the assign- 

 ment of weights is very likely to change during the carrying out 

 of the comparisons. Thus one will come at the end to give large 

 weight to differences which at first were marked as small, or 

 vice versa. This may be avoided very largely by making up 

 the comparison table in irregular order. Thus, instead of com- 

 paring plant number 1 with all the others, then plant number 2, 

 etc., as in the usual method, one takes first the comparisons of 

 plant number 1 than, for instance, those of plant number 80 

 then those of plant number 36 and so on, skipping from one part 

 of the table to another until all is finally complete. In this way 

 any undetected changes of mental criteria during the course of 

 the comparison will be distributed over the entire series and 

 without important effect on the results. 



